Our data point to a force apart from thermodynamic stability that conserves buried and interface residues in H3. This driving force of conservation found in buried H3 residues has not been observed or characterized in buried residues of other highly conserved proteins to our knowledge. Statistically significant correlation between Medusa derived sequence entropies and evolutionary entropies has been shown before in other proteins 
, and in H4 in this study, implying stability as a driving force for conservation buried residues. Further, when we performed similar analysis on all buried positions of actin (PDB ID 1J6Z) and tubulin (PDB ID 1Z5V), we observe statistically significant correlation between evolutionary and Medusa entropies (). Thus, there is correlation between thermodynamic stability and evolutionary conservation in the buried residues found in other highly conserved proteins, namely actin and tubulin. These results reveal H3 as the only protein known so far, whose conservation of core residues is not driven by stability alone. We therefore suggest there is a novel function associated with the buried and interface residues of H3 that is driving the unique level of conservation.
Correlation between evolutionary entropy and Medusa entropy in highly conserved proteins.
What might be the function of the conserved buried residues in H3? Given these residues are not found to be post-translationally modified, and their strict conservation is independent of H3-H3′ and H3-H4 stability, it could be suggested that they may be playing a role in histone chaperone interactions and deposition. Asf1 is one such histone chaperone that facilitates the deposition of histones in chromatin during replication. Even though Asf1 binds and/or competes for the H3-H4 dimer by interacting with residues in the H3-H3′ interface 
, it cannot account for the conservation of the residues it interacts with in H3, as the interacting residues from Asf1 are not similarly conserved. In addition, Asf1 is not known to interact with the H3 buried residues we examined. We do not rule out the possibility that other histone chaperones or other histone interacting proteins interact with the buried and interface residues in H3/H4 for a functional purpose, but it is striking to note that even though most of these proteins may also interact with H4, conservation of H4 can be accounted for by thermodynamic stability alone.
Another possibility for the increased conservation of buried residues of H3 could be the need to tightly regulate and fine-tune the stability of nucleosomes during transcriptional regulation, as a slight increase or decrease of nucleosome stability could have profound effects on cellular processes like transcription. We find evidence for this hypothesis in the observation that H3 variants such as H3.3 function by modulating nucleosome stability 
. Although the core of H3.3 differs from canonical H3 in humans at just three positions, the destabilization of H3.3 containing nucleosomes has been shown to be important in transcriptional regulation 
. Thus, major changes in cellular function due to minor perturbations in H3 sequence suggest the need for tight control of nucleosome stability. Such control may explain why the core residues of H3 are so highly conserved.
We conclude that an unknown set of factors is driving conservation of H3 to a degree that has not been found in any other protein to date. The significance of these residues outside of histone fold interactions awaits further discovery. Our finding of an unexpected level of sequence conservation, not demonstrated before in a protein to our knowledge, suggests the ability to predict functional roles of amino acid residues apart from imparting thermodynamic stability to a given protein.